Math 2280-001 Fri Feb 6

advertisement
Math 2280-001
Fri Feb 6
, Let's start with Exercise 7 from Wednesday's notes. This will also be an opportunity to summarize the
three numerical methods for solving DE's that we discussed: i.e. Euler, Improved Euler, and Runge Kutta,
and for you to ask any questions you might have related to the 2.4-2.6 material. There is an application
problem in this upcoming week's homework that will use numerical methods.
................................................................................................................
, Then, begin Chapter 3, which is about higher order linear differential equations and applications.
3.1 Second order linear differential equations, and vector space theory connections to Math 2270:
Definition: A vector space is a collection of objects together with an "addition" operation "+", and a "scalar
multiplication" operation, so that the rules below all hold.
(α) Whenever f, g 2 V then f C g 2 V . (closure with respect to addition)
(β) Whenever f 2 V and c 2 =, then c$f 2 V. (closure with respect to scalar
multiplication)
As well as:
(a) f C g = g C f (commutative property)
(b) f C g C h = f C g C h (associative property)
(c) d 0 2 V so that f C 0 = f is always true.
(d) c f 2 V dKf 2 V so that f C Kf = 0 (additive inverses)
(e) c$ f C g = c$f C c$g (scalar multiplication distributes over vector addition)
(f) c1 C c2 $f = c1 $f C c2 $f (scalar addition distributes over scalar multiplication)
(g) c1 $ c2 $f = c1 c2 $f (associative property)
(h) 1$f = f, K1 $f =Kf, 0$f = 0 (these last two actually follow from the others).
Examples you've seen in Math 2270:
(1) =m, with the usual vector addition and scalar multiplication, defined component-wise
(2) subspaces W of =m, which satisfy (α),(β), and therefore automatically satisfy (a)-(h), because the
vectors in W also lie in =m.
Maybe you've also seen ...
Exercise 1) In Chapter 3 we focus on the vector space
V = C = d f : =/= s.t. f is a continuous function
and its subspaces. Verify that the vector space axioms for linear combinations are satisfied for this space
of functions. Recall that the function f C g is defined by f C g x d f x C g x and the scalar
multiple c f x is defined by c f x d c f x . What is the zero vector for functions?
Recall that the vector space axioms are exactly the arithmetic rules we use to work with linear combination
equations. In particular the following concepts are defined in any vector space V.
, the span of a finite collection of functions f1 , f2 , ... fn .
, linear independence/dependence for a collection of functions f1 , f2 , ... fn .
, subspaces of V
, bases and dimension for finite dimensional subspaces. (The function space V itself is infinite
dimensional, meaning that no finite collection of functions spans it.)
Definition: A second order linear differential equation for a function y x is a differential equation that
can be written in the form
A x y##C B x y# C C x y = F x .
We search for solution functions y x defined on some specified interval I of the form a ! x ! b, or
a, N , KN, a or (usually) the entire real line KN,N . In this chapter we assume the function A x
s 0 on I, and divide by it in order to rewrite the differential equation in the standard form
y##C p x y#C q x y = f x .
Definition: The DE above is called homogeneous if the right hand side f x is the zero function,
f x h 0. If f is not the zero function, the DE is called nonhomogeneous (or inhomogeneous).
One reason the DE above is called linear is that the "operator" L defined by
L y d y##C p x y#C q x y
satisfies the so-called linearity properties
(1) L y1 C y2 = L y1 C L y2
(2) L c y = c L y , c 2 = .
(Recall that the matrix multiplication function L x d A x satisfies the analogous properties. Any time we
have have a transformation L satisfying (1),(2), we say it is a linear transformation. )
Exercise 2a) Check the linearity properties (1),(2) for the differential operator
L y d y##C p x y#C q x y .
2b) Use these linearity properties to show that
Theorem 0 the solution space to the homogeneous second order linear DE
y##C p x y#C q x y = 0
is closed under addition and scalar multiplication, i.e. it is a subspace. Notice that this is the "same" proof
one uses to show that the solution space to a homogeneous matrix equation A x = 0 is a subspace.
Exercise 3) As an example, find the solution space to the following homogeneous differential equation for
y x
y##C 2 y#= 0
on the xKinterval KN! x !N. Notice that the solution space is the span of two functions. Hint: This is
really a first order DE for v = y#.
Exercise 4) Use the linearity properties to show
Theorem 1 All solutions to the nonhomogeneous second order linear DE
y##C p x y#C q x y = f x
are of the form y = yP C yH where yP is any single particular solution and yH is some solution to the
homogeneous DE. (yH is called yc, for complementary solution, in the text). Thus, if you can find a single
particular solution to the nonhomogeneous DE, and all solutions to the homogeneous DE, you've actually
found all solutions to the nonhomogeneous DE.
Theorem 2 (Existence-Uniqueness Theorem): Let p x , q x , f x be specified continuous functions on
the interval I , and let x0 2 I . Then there is a unique solution y x to the initial value problem
y##C p x y#C q x y = f x
y x0 = b0
y# x0 = b1
and y x exists and is twice continuously differentiable on the entire interval I .
Exercise 5) Verify Theorems 1 and 2 for the interval I = KN,N and the IVP
y##C 2 y#= 3
y 0 = b0
y# 0 = b1
Unlike in the previous example, and unlike what was true for the first order linear differential equation
y#C p x y = q x
there is not a clever integrating factor formula that will always work to find the general solution of the
second order linear differential equation
y##C p x y#C q x y = f x .
Rather, we will usually resort to vector space theory and algorithms based on clever guessing to solve
these differential equations. It will help to know
Theorem 3: The solution space to the second order homogeneous linear differential equation
y##C p x y#C q x y = 0
is 2-dimensional.
This Theorem is illustrated in Exercise 2 that we completed earlier. Theorem 3 and the techniques we'll
actually be using going forward are illustrated by
Exercise 6) Consider the homogeneous linear DE for y x
y##K2 y#K3 y = 0
6a) Find two exponential functions y1 x = er x , y2 x = er x that solve this DE. Deduce that arbitrary
linear combinations of y1 , y2 also solve the DE.
6b) Show that every IVP
y##K2 y#K3 y = 0
y 0 = b0
y# 0 = b1
can be solved with a unique linear combination y x = c1 y1 x C c2 y2 x .
6c) Use your work from part b to explain why the solution space is two-dimensional.
6d) Now consider the nonhomogeneous DE
y##K2 y#K3 y = 9
Notice that yP x =K3 is a particular solution. Use this information and superposition (linearity) to find
the solution to the initial value problem
y##K2 y#K3 y = 9
y 0 =6
y# 0 =K2.
Download